The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references documentation and official service URLs from well-known technology providers including Cloudflare (PartyKit), Supabase, Liveblocks, and Ably. These references are used for architectural decision-making and are documented neutrally as legitimate technology resources.
[COMMAND_EXECUTION]: The skill includes two Python scripts (check_ws_smoke.py and ws_smoke_test.py) for validating WebSocket endpoint connectivity. These scripts use standard networking libraries such as socket, ssl, and websockets. Analysis of the code confirms they do not perform unauthorized system modifications, sensitive file access, or remote code execution from untrusted sources.
[DATA_EXPOSURE]: There are no hardcoded credentials, API keys, or attempts to access sensitive system paths (e.g., SSH keys, environment files) within the skill instructions or scripts.
[PROMPT_INJECTION]: The skill body consists of technical guidance and best practices. It does not contain instructions aimed at overriding AI safety filters, system prompt extraction, or manipulating the agent's core behavioral constraints.
[SAFE]: The skill follows secure management practices by recommending externalized secret management and ephemeral storage for transient presence data.

software-realtime